On Thu, Oct 13, 2011 at 1:24 PM, Kevin Pouget <kevin.pouget@gmail.com> wrote:
> On Wed, Oct 12, 2011 at 7:06 PM, Tom Tromey <tromey@redhat.com> wrote:
>>>>>>> "Pedro" == Pedro Alves <pedro@codesourcery.com> writes:
>>
>> Pedro> We should definitely reimplement these enums in a table in a
>> Pedro> .def file.  That's 2 places already that map the enums to
>> Pedro> something else.  This one is missing UNWIND_UNAVAILABLE.
>>
>> Makes sense to me.
>>
>> Pedro> And what do you think of making UNWIND_FIRST_ERROR
>> Pedro> an alias like in my patch?  Do you think that's likely
>> Pedro> to break anything?
>>
>> I think it would be fine.
>
>
> Hello,
>
> here is a patch which implements the enums in a .def file, as
> suggested earlier. Please let me know what you think about it (tested
> with no regression on X86_64)
>
> UNWIND_FIRST_ERROR is now an "enum alias", instead of being defined by
> the preprocessor (the only way I managed to implement it), as well as
> UNWIND_FIRST and UNWIND_LAST, used for bound-checking in py-frame.c
>
> I also moved the original enum comments to the .def file with no modification
>
>
> Thanks,
>
> Kevin

(with the file attached this time, sorry for the repost)